Ray casting vs frustum culling
In the classic scenario of a large grid of quads making up a terrain, what would you suggest for culling out non-visibile tiles, raycasting from the camera point of view and checking intersection with each tile vertex, or culling out the quad using the frustum? Before implementing one or the other method I'd like to understand which method is the most computationally intensive.
Thanks
- a
neither. Since you're using Quad shaped tiles, store the tiles as a quadtree, and then use frustum culling with the quadtree structure. linky, another link
Right, I intended to use the recursive structure of a quadtree. But isn't that actually an abstraction that would apply to raycasting too?
Let's say I have a 128x128 tiles terrain. First I check each of the 4 64x64 "macrotiles" for visiblity. For each one which passes the test, I split it into 4 16x16 macrotiles and check it... and so on recursively.
The check can still be performed either by raycasting or frustum culling.
Let's say I have a 128x128 tiles terrain. First I check each of the 4 64x64 "macrotiles" for visiblity. For each one which passes the test, I split it into 4 16x16 macrotiles and check it... and so on recursively.
The check can still be performed either by raycasting or frustum culling.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement